The invention relates to optical imaging arrangements used in exposure processes, in particular to optical imaging arrangements of microlithography systems. It further relates to a method of supporting components of an optical projection unit. The invention may be used in the context of photolithography processes for fabricating microelectronic devices, in particular semiconductor devices, or in the context of fabricating devices, such as masks or reticles, used during such photolithography processes.
Typically, the optical systems used in the context of fabricating microelectronic devices such as semiconductor devices comprise a plurality of optical element units comprising optical elements, such as lenses and mirrors etc., arranged in the light path of the optical system. Those optical elements usually cooperate in an exposure process to transfer an image of a pattern formed on a mask, reticle or the like onto a substrate such as a wafer. The optical elements are usually combined in one or more functionally distinct optical element groups. These distinct optical element groups may be held by distinct optical exposure units. In particular with mainly refractive systems, such optical exposure units are often built from a stack of optical element modules holding one or more optical elements. These optical element modules usually comprise an external generally ring shaped support device supporting one or more optical element holders each, in turn, holding an optical element.
Optical element groups comprising at least mainly refractive optical elements, such as lenses, mostly have a straight common axis of symmetry of the optical elements usually referred to as the optical axis. Moreover, the optical exposure units holding such optical element groups often have an elongated substantially tubular design due to which they are typically also referred to as lens barrels.
Due to the ongoing miniaturization of semiconductor devices there is a permanent need for enhanced resolution of the optical systems used for fabricating those semiconductor devices. This need for enhanced resolution obviously pushes the need for an increased numerical aperture (NA) and increased imaging accuracy of the optical system.
One approach to achieve enhanced resolution is to reduce the wavelength of the light used in the exposure process. In the recent years, approaches have been made to use light in the extreme ultraviolet (EUV) range using wavelengths ranging from 5 nm to 20 nm, typically about 13 nm. In this EUV range it is not possible to use common refractive optics any more. This is due to the fact that, in this EUV range, the materials commonly used for refractive optical elements show a degree of absorption that is too high for obtaining high quality exposure results. Thus, in the EUV range, reflective systems comprising reflective elements such as mirrors or the like are used in the exposure process to transfer the image of the pattern formed on the mask onto the substrate, e.g. the wafer.
The transition to the use of high numerical aperture (e.g. NA>0.4 to 0.5) reflective systems in the EUV range leads to considerable challenges with respect to the design of the optical imaging arrangement.
One of the crucial accuracy requirements is the accuracy of the position of the image on the substrate, which is also referred to as the line of sight (LoS) accuracy. The line of sight accuracy typically scales to approximately the inverse of the numerical aperture. Hence, the line of sight accuracy is a factor of 1.4 smaller for an optical imaging arrangement with a numerical aperture NA=0.45 than that of an optical imaging arrangement with a numerical aperture of NA=0.33. Typically, the line of sight accuracy ranges below 0.5 nm for a numerical aperture of NA=0.45. If double patterning is also to be allowed for in the exposure process, then the accuracy would typically have to be reduced by a further factor of 1.4. Hence, in this case, the line of sight accuracy would range even below 0.3 nm.
Among others, the above leads to very strict requirements with respect to the relative position between the components participating in the exposure process. Furthermore, to reliably obtain high-quality semiconductor devices it is not only necessary to provide an optical system showing a high degree of imaging accuracy. It is also necessary to maintain such a high degree of accuracy throughout the entire exposure process and over the lifetime of the system. As a consequence, the optical imaging arrangement components, i.e. the mask, the optical elements and the wafer, for example, cooperating in the exposure process must be supported in a defined manner in order to maintain a predetermined spatial relationship between the optical imaging arrangement components as well to provide a high quality exposure process.
To maintain the predetermined spatial relationship between the optical imaging arrangement components throughout the entire exposure process, even under the influence of vibrations introduced, among others, via the ground structure supporting the arrangement and/or via internal sources of vibration disturbances, such as accelerated masses (e.g. moving components, turbulent fluid streams, etc.), the optical imaging arrangement as well as the under the influence of thermally induced position alterations, it is necessary to at least intermittently capture the spatial relationship between certain components of the optical imaging arrangement and to adjust the position of at least one of the components of the optical imaging arrangement as a function of the result of this capturing process.
In conventional systems, this process of capturing the spatial relationship between components cooperating in the exposure process is done via a metrology system using a central support structure for the optical projection system and the substrate system as a common reference in order to be able to readily synchronize motion of the actively adjusted parts of the imaging arrangement.
On the other hand, an increase in the numerical aperture, typically, leads to an increased size of the optical elements used, also referred to as the optical footprint of the optical elements. The increased optical footprint of the optical elements used has a negative impact on their dynamic properties and the control system used to achieve the above adjustments. Furthermore, the increased optical footprint typically leads to larger light ray incidence angles. However, at such increased larger light ray incidence angles transmissivity of the multi-layer coatings typically used for generating the reflective surface of the optical elements is drastically reduced, obviously leading to an undesired loss in light power and an increased heating of the optical elements due to absorption. As a consequence, even larger optical elements have to be used in order to enable such imaging at a commercially acceptable scale. These circumstances lead to optical imaging arrangements with comparatively large optical elements having an optical footprint of up to 1 m×1 m and which are arranged very close to each other with mutual distances ranging down to less than 60 mm.
Several problems result from this situation. First, irrespective of the so-called aspect ratio (i.e. the thickness to diameter ratio) of the optical element, a large optical element generally exhibits low resonant frequencies. While, for example, a mirror with an optical footprint of 150 mm (in diameter) and a thickness of 25 mm typically has resonant frequencies above 4000 Hz, a mirror with an optical footprint of 700 mm, typically, hardly reach resonant frequencies above 1500 Hz even at a thickness of 200 mm. Furthermore, increased size and weight of the optical elements also means increased static deformation due to variations of the gravitational constant at different locations all over the world, which impairs imaging performance when uncorrected.
With conventional support systems striving to support the optical elements at a maximum rigidity (i.e. at maximized resonant frequencies of the support system) low resonant frequencies of the optical element itself lead to a reduction of the adjustment control bandwidth and, hence, reduced position accuracy.
Furthermore, large optical elements resulting in a large object to image shift ultimately lead to a large and less rigid support structure for the optical system. Such a less rigid support structure not only contributes to further restrictions of adjustment control performance, but also residual errors due to quasi-static deformations of the structure caused by residual low frequency vibration disturbances. Hence, the negative effects of vibration disturbances become even more prominent.
Finally, the increased thermal load on the optical elements used (due to light energy absorption) and the increased throughput desired for such systems requires increased cooling efforts, in particular, higher flow rates of the cooling fluids used. This increased cooling flow rate is prone to lead to an increase in the vibration disturbances introduced into the system, in turn leading to reduced line-of-sight accuracy.